LT8310 Duty Control + Minimum Load & FB MODE

Hello,

I am using the LT8310 in a design for a 9V to 15V, 60W output forward converter. I do not need precise line regulation and am pleased with the performance form the duty mode control alone. 

That being said -- I do have operational constraints to supply a very light load on occasion and remain under ~14 to 15V (but the output drifts up to 18 to 20V or higher for light loads). The nominal load expected is above 10W and typically more in the 30W range -- however it could feasibly drop down to 1W or even much less.

One solution is to add a passive or active load... but that uses unnecessary power. Another solution is to add isolated feedback to allow for current mode control pulse skipping -- but I do not want to have to compensate the isolated feedback network and worry about loop stability (I am pleased with the simple CDFILT compensation in duty mode control).

While experimenting with isolated feedback (using an ADuM4190) in LTSpice I found something interesting... If I set RVDIN RSET resistor to get 12V in duty mode control... But then I add isolated feedback with a setpoint of 14.5V I get the following behavoir:

• When load is above Imin (>0.5A) --> Duty mode control appears to dominate and the converter stays at ~12V
• When load is below Imin --> Vout drifts upwards to ~14.5V and then the feedback control appears to kick in and regulate the output

I am a fan on this implementation because it allows me to not worry so much about closed loop stability since the light load condition does not happen often / is not a critical use case.

I am concerned however that this behavoir is not "real" / is something wrong with the SPICE model for the part. 

The datasheet states that when FBX is above 0.3V -- current mode control is in place, not duty mode... however per the simulation it appears that if FBX > 0.3V, duty mode control takes precedence so long as Imin load is satisfied for the given RDVIN RSET setpoint and the output voltage is less than the feedback setpoint. Unless I missed it... this behavior is not discussed in the datasheet.

So my question is essentially: Is the simulation correct? Or is it just a some incorrect model setting for the part?

Relevant circuitry:

Waveforms showing what appears to be current mode control taking over at light load and duty mode control taking over once load is large enough:

Versus the behavior with FBX shorted to GND and no isolated feedback connected to Vc: